The long-term objectives of this research are to understand the psychosocial, behavioral and biomedical determinants of differential trajectories of physical and cognitive function in the elderly. We are especially interested in identifying protective factors that promote the retention of higher levels of cognitive or physical function into very old age or those that promote resilience or ability to recover after significant medical events. We aim to accomplish this by conducting an 11 year follow-up survey of the New Haven EPESE, a community-based epidemiologic cohort study of 2812 men and women who were living, non- institutionalized in New Haven, CT in 1982. The cohort has been under continual surveillance for hospitalization, nursing home admission, and mortality since 1982. Subjects have had in-depth face-to-face assessments of physical and cognitive function in 1982, 1985, and 1988 with briefer telephone contacts in interim years. 1457 subjects were alive at the end of 1991. Approximately 1250 will be available to participate in a fourth face-to-face assessment. We propose to reinterview and re-examine subjects in 1993-1994, conduct surveillance in all 3 years and analyses in Years 02 and 03. The specific aims are: 1. To determine and describe the cognitive and physical abilities of a population-based cohort of men and women all of whom will be at least 75 years of age in 1993 and for whom there will be 11 years of data on longitudinal changes in function. 2. To test hypotheses concerning the role of socioeconomic status, social networks and support, social participation, health practices such as smoking, alcohol consumption and physical activity and biomedical risk factors such as obesity, high blood pressure and diabetes in predicting the retention or decline in physical and cognitive function. 3. To obtain detailed clinical information on three acute-onset medical events: MI, stroke, and hip fracture and investigate the factors that predict long-term course of cognitive and physical function and risk of institutionalization controlling for detailed clinical conditions related to prognosis. The strengths of this study are 1)its heterogeneous population-based sample with high response rates (between 90-97%) on all follow-up surveys and less than 1% of subjects lost to follow-up; 2)a decade of rich longitudinal data on both risk and protective factors and outcomes; 3) multiple hospital and long-term care surveillance systems which identify and verify over 95% of such events; 4)a cohort which is now entering its period of highest risk.